Stat-spotting: a field guide to identifying dubious data

Getting started

Spotting questionable numbers

Background

Statistical benchmarks

Severity and frequency

Varieties of dubious data

Blunders

The slippery decimal point

Botched translations

Misleading graphs

Careless calculations

Sources: who counted

and why?

Big round numbers

Hyperbole

Shocking claims

Naming the problem

Definitions: what did they count?

Broad definitions

Expanding definitions

Changing definitions

The uncounted

Measurements: how did they count?

Creating measures

Odd units of analysis

Loaded questions

Raising the bar

Technical measures

Packaging: what are they telling us?

Impressive formats

Misleading samples

Convenient time frames

Peculiar percentages

Selective comparisons

Statistical milestones

Averages

Epidemics

Correlations

Discoveries

Debates: what if they disagree?

Causality debates

Equality debates

Policy debates

Stat-spotting on your own

Summary: common signs of dubious data

Better data: some characteristics

Afterword: if you had no idea things were that bad, they probably aren't

Suggestions for those who want to continue stat-spotting.